Skip to content
Will AI Redefine Programmers? Four Hard Questions from Tencent's Developer Pioneers
Industry Insights 14 min read

Will AI Redefine Programmers? Four Hard Questions from Tencent's Developer Pioneers

Tencent Cloud Developer Pioneers share four core anxieties about AI's impact on programming: shifting bottlenecks from implementation to judgment, human value in taste and stories, learning through experimentation over knowledge hoarding, and programmers evolving from module owners to independent product creators.

Tencent Tech
Tencent Tech
Tencent Tech

Question 1: What Has AI Actually Changed?

The initial excitement was simple: AI handles coding, documentation, environment setup, and repetitive tasks. Meimaobing (TDP) moved OpenClaw into CNB and turned it into a repository NPC engineer that reads code, checks docs, and replies to comments when mentioned in issues. He also built a "snarky review" tool that scans a user's repos, PRs, and contributions to generate a data-backed roast. Tasks once done manually are now delegated to agents.

But as AI starts doing real work, new problems surface: large-model judgments lack verifiable sources and stable reproducibility; agent output still requires human line-by-line verification. Completing a task does not mean the result is reliable. Xuan (TDP) lets agents understand software UIs via screenshots and OCR, even deciding when to speak in QQ groups. His core insight: "Models are evolving; the real engineering challenge is making them survive stably in real environments."

The fundamental contradiction: implementation cost has plummeted, so ideas become runnable programs in minutes. Yet software value doesn't grow with code volume. Agents can generate ten solutions in minutes; developers must still judge which is worth adopting, which parts to keep. AI can rapidly create technical debt and new problems that consume hours of iteration. Execution gets easier, but expectations rise — half-hour AI generation replaces day-long manual work, and the saved time fills with more demands and complexity.

After a year, TDP members see the bottleneck shifting from "how to implement" to "what to implement, how to verify, how to deliver." AI doesn't eliminate engineering problems; it pushes developers into positions requiring more judgment and accountability.

Question 2: Where Is Human Value?

Programmers traditionally prove worth by solving hard technical puzzles. When AI does more of that, we ask: what remains to be proud of? Hunsheng (TDP) developed FontGate handling 546 fonts; first launch took 117 seconds. Through iterative collaboration with AI, he compressed it to 2.6 seconds. The experience upgrade came not from development speed but from a developer's judgment on product experience.

AI can implement features and suggest optimizations, but humans must decide: what problem is worth solving? What counts as "done well"? Why spend time here instead of chasing the next hot tool? Hunsheng says: "AI lowers the barrier from zero to one; it does not lower the barrier of judgment." Implementation difficulty should not be the sole measure of value. In the AI era, coders create value by knowing "why I write" and "what counts as writing well."

Ar-Sr-Na (TDP) keeps 60,000+ personal photos. He notes: "What AI cannot take away is your story." Even if future models simulate identical images, they cannot reconstruct the memories and stories behind those photos. The interim answer: AI multiplies digital construction efficiency, but humans must still make sensibility-driven judgments. "Judgment rights are not outsourced — that's the bottom line." This isn't opposition to AI; it's a reaffirmation of programmer value under a new division of labor.

Question 3: Does Learning Still Have Value?

2026's frustration: "can't keep up." Just as you master a tool, a new framework appears; a workflow is built, and the next agent automates it. We used to believe practice makes perfect and experience accumulates. Now we doubt: if today's learning becomes a button in months, what's the point?

Hunsheng (TDP), a self-taught dropout who once bought domains, servers, and racked machines, now migrates dev, build, and workspace to CNB, shedding self-maintained infra. Did he waste effort learning the old way? Precisely because he understands the principles and costs, he can better judge what to do himself and what to delegate to platforms.

Anye (TDP) started as a "666" viewer in TDP livestreams. Leveraging AI and the TDP community, he advanced rapidly and now shares AI-collaboration experience at offline events. His growth came not from mastering every technology but from constant experimentation and turning practice into reusable insights for others. TDP consensus: "The group isn't comparing who knows more; it's comparing who tries more."

AI-era learning shifts from solo knowledge hoarding to a cycle: ask questions, use tools to experiment, verify results, correct understanding, exchange experiences with peers. That's why genuine technical exchange between people becomes more precious.

Question 4: What Will Programmers Be Redefined As?

At 2025's 1024, Xiong (TDP) judged: AI isn't the threat; the gap is between those who use AI and those who don't. A year later, he questions that premise. If every programmer uses AI? If AI no longer needs a "better user"? How much advantage does "using AI better than others" confer when tools commoditize and efficiency edges vanish in days or seconds?

Statically, the "programmer" role is being rewritten. Traditional tasks — basic scripting, repetitive work, simple bug fixes — AI already does better; refactoring, DB design, complex debugging — AI+human excel. But flip the view: programmers built tools that share their own labor, gaining unprecedented creative capacity. Previously a cog in a large system; now, with AI, one person can take an idea from requirement through design, development, test, and deployment.

Xiong sees another possibility: programmers shifting from system cogs to independent product creators. Not everyone must start a business, nor can one person solve everything. The key: programmers can define work beyond lines of code — participating in problem definition, product design, final delivery, and attempting ideas once impossible due to resource constraints.

TDP members conclude: instead of betting on a skill machines will never touch, continuously expand the problems you can solve and the value you can create. Don't just be an AI user; become someone who harnesses AI to build genuinely useful products. The answer isn't unique, but it's always human-given.

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

developer anxietyAI agents in developmentAI impact on programminghuman judgment vs AIlearning in AI eraprogrammer career redefinitionsoftware engineering bottlenecksTencent Cloud TDP
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.